In this set of short exercises, we will continue with our data wrangling tasks.

As per usual, before we can do anything else, we first need to load the tidyverse package(s) and import the data.

library(tidyverse)

gpc <- read_csv("./data/ZA5667_v1-0-0_Stata14_synthetic-data.csv")

In the first exercises, we want to work with the the variables that assess how much people trust specific people or institutions in dealing with the Corona virus. Remember that in the process of creating the synthetic data we are using here, all values < 0 have been set to NA. However, if we look into the codebook for the original data set, we see that the trust items contain the value 98, representing the response option “I don’t know”. For our analyses, we probably want to treat this as a missing value.

1

Overwrite the variable measuring trust in the federal government in the gpc dataframe, so that 98 is a missing value.

For recoding a value as NA in a particular variable, we need to combine two functions from dplyr: One that creates a new variable or overwrites an existing one and another one that converts specific values to NA.The variable that we want to do this for is named hzcy048a.

solution 

gpc <- gpc %>% 
  mutate(hzcy048a = na_if(hzcy048a, 98))

2

Now that we have recoded 98 as NA in this variable, let´s create a new variable named distrust_gov that captures distrust instead of trust in the federal government.

Remember that the correct syntax for recoding values with the corresponding dplyr function is old value (enclosed in backticks) = new value.

solution 

gpc <- gpc %>% 
  mutate(distrust_gov = recode(hzcy048a,
                               `5` = 1,
                               `4` = 2,
                               `2` = 4,
                               `1` = 5))

3

How many of the simulated respondents do not have a missing value for the variable political_orientation? To answer this question, please use a function from the tidyr package that allows you to exclude cases with missing values.

To count the number of cases, you can use the base R function nrow() at the end of your pipe.

solution 

gpc %>% 
  drop_na(political_orientation) %>% 
  nrow()
## [1] 3706

4

Create a new binary variable called married in the gpc dataframe that has the value 1 if the individual is married and 0 if not.

You can make use of the function ifelse(). The variable we need to use here is called marstat where the value 1 indicates that the person is married.

solution 

gpc <- gpc %>% 
  mutate(married = ifelse(marstat == 1, 1, 0))

5

Let’s create another new variable. This time, it should be a character variable named age_3cat that has the following unique values representing the respective age categories: “up to 40 years”, “41 to 60 years” and “older than 60 years”.

You can use the between() helper function in combination with the dplyr function for conditional variable creation/transformation. The required existing variable is called age_cat. As a side note: For your actual data, you probably would want to have such a variable as an ordered factor, but we have not covered factors in this workshop, so we’ll stick with a simple character variable here.

solution 

gpc <- gpc %>% 
  mutate(age_3cat = case_when(
    between(age_cat, 1, 4) ~ "up to 40 years",
    between(age_cat, 5, 7) ~ "41 to 60 years",
    age_cat > 7 ~ "older than 60 years"
    ))

6

As a final data wrangling exercise, let’s create a variable called info_sum that represents the sum of all information sources that the (simulated) respondents reported to use to get current information about the Corona virus. You can find the names of the variables we need for this in the previous set of exercises or in the codebook for the original data set. However, here we do not want to include the last variable in this series (hzcy095a) as that one indicates whether people stated that they do not inform themselves at all.

Remember that we need to perform this operation rowwise. We can also use a special version of the across() helper function that allows us to combine values from multiple columns. If you can’t or don’t want to consult the codebook for the original data, the variable names we need for this task range from hzcy084a to hzcy093a.

solution 

gpc <- gpc %>% 
  rowwise() %>%
  mutate(info_sum = sum(c_across(hzcy084a:hzcy093a),
                            na.rm = TRUE)) %>% 
  ungroup()